Infrastructure planning, fleet sizing, and scheduling for E-public transport

Abstract In this work, we analyze the integrated problem of sizing and locating charging infrastructure as well as sizing a fleet of traditional and electric buses to cover a set of scheduled trips. The available buses include the following: traditional diesel vehicles (TV), standard electric vehicles (EV), and electric vehicles equipped with ultra-fast recharging (UV). The latter exploit the latest available technology, and therefore, their purchasing cost and the cost of their compatible charging infrastructure are very high. However, their recharging time, which is 12 times less than that of EV, makes them useful when dealing with a tight schedule. This study aims to determine how many EV and UV should be purchased and how many chargers of each type should be installed in each recharging station. Also, it aims to outline a feasible schedule for each bus, taking into consideration the eventual stops for recharging, in order to minimize the overall cost. To achieve the study’s aims, we create an innovative and elegant integer programming formulation in which we model multiple visits to recharging stations and propose an exact method based on Combinatorial Benders Cuts, in which the master problem is modeled as a temporal bin packing problem with side constraints. In this computational study, we determine the conditions in which the exploitation of ultra-fast technology is useful and analyze the advantages of exploiting a mixed fleet.